The Verbose Context Problem in Medical Records

2026-06-28

AI Summary

This study addresses the inefficiency caused by excessively long clinical narratives in electronic health records—particularly in population-scale longitudinal analyses, where individual patient records often exceed 400,000 tokens. We formally define and isolate the “verbose context problem” specific to medical settings. To this end, we introduce PopMedQA, a benchmark for evaluating model reasoning over longitudinal population-level medical records, and develop the neopatient library to generate controllable synthetic patient histories. Using these resources, we systematically assess prompting strategies, context compression techniques, and agent-driven task decomposition. Our experiments reveal that general-purpose approaches yield limited gains, whereas input optimization incorporating medical structural priors substantially improves performance, thereby validating the efficacy of domain-specific benchmarks and synthetic data frameworks for scalable medical reasoning.

Abstract

The verbose context problem occurs when structured concepts have token-inefficient textual representations. This bottleneck is acute in population health: cohort-level analysis of longitudinal patient records requires reasoning over thousands of medically-coded events, often exceeding 400K tokens in total. We present PopMedQA, a benchmark isolating this problem through computational tasks on groups of longitudinal patient records. We construct the benchmark using neopatient, a new library for language-controlled generation of artificial patient records. Through extensive ablations -- including prompting strategies, prompt compression, and agentic decomposition -- we find that domain-independent methods fail to alleviate the verbose context problem. There remains significant opportunity to exploit domain-specific structure in language model inputs for population-scale reasoning.

Problem

• verbose context problem
• medical records
• population health
• longitudinal patient records
• token inefficiency

Innovation

• verbose context problem
• PopMedQA
• neopatient
• longitudinal patient records
• domain-specific structure